Introducing the real driving emission(RDE)test procedures into the light-duty vehicle emission regulations is an effective technical solution to control vehicle pollutant emissions and improve the quality of the air environment.The China-6 Emission Regulation has added the requirements for RDE tests.However,the current RDE regulation has not included the vehicles’cold-start emissions into the regulatory scope.To reflect the emission level of light-duty vehicles more comprehensively and realistically,it is necessary to consider the cold-start emissions of vehicles in RDE tests.To investigate the evaluation method of RDE tests for light-duty vehicles with considering cold-start emissions,four light-duty gasoline vehicles meeting the requirements of China-6a emission standards were used to perform a series of RDE tests on four selected test routes.The specific research contents are as follows:To evaluate the contribution of cold-start emissions to the urban and total trips,a total of eight cold-start RDE tests were completed using one of the gasoline vehicles on the four test routes.The results show that the average contributions of CO emissions from cold-start to urban and total trip CO emissions were 28.3%and 11.3%respectively.The average contributions of NOx emission from cold-start were 31.9%and 14.4%,and the average contributions of PN emission from cold-start were 39.8%and 5.8%.Compared with the CO,NOx and PN emissions,the contributions of cold-start CO2 emissions to urban and total trip emissions were relatively low,which were 5.8%and 2.6%respectively.Thus,it is necessary to consider cold-start emissions in the RDE tests.Due to the complexity of the RDE test boundary conditions(such as environmental conditions,driving routes,vehicle operating conditions,etc.),the cold-start test results show poor repeatability and large variations.Considering cold-start emissions can make the problem of poor reproducibility of RDE tests results even more pronounced.Accordingly,this paper further explores the reasons for the differences in cold-start emissions between the eight tests.In the full cold-start trip,CO and CO2 emissions are significantly correlated with average coolant temperature,while NOx and PN emissions are strongly correlated with average vehicle speed.Additionally,the cold-start emission results under RDE tests are quite different,which is closely related to the dynamic parameter v·apos(velocity times positive acceleration)in the first 50 seconds of cold-start.It is recommended that the dynamic parameters at the initial stage of cold-start(e.g.,the first 50 seconds of cold-start)should be strictly limited when amending the RDE test procedure considering cold-start emissions.In addition,the moving average window method(MAW,adopted by the current China-6 Emission Regulation)and cumulative average method(CA,adopted by the current Euro-6 Emission Regulation)were used to calculate and analyze the RDE test data from three light-duty gasoline vehicles,so as to explore the emission evaluation method suitable for the RDE test considering cold-start.A comparative analysis of the calculation results shows that both the methods are flawed when processing RDE data including cold-starts.Specifically,the MAW method unreasonably underestimates cold-start emission data in the urban and total trip emission evaluation.The CA method tends to increase the inconsistency of the RDE evaluation results especially after the cold-start emissions are considered in the final evaluation results.Considering the inherently different characteristics between the cold-start emissions and non-cold-start emissions,it is proposed that the RDE test trip should be divided into a cold-start trip and a non-cold-start trip,and use the CA method and MAW method to evaluate emissions of the two trips respectively.Furthermore,the pollutant emission factors in the cold-start trip are corrected referring to the CO2 emissions emitted during the cold-start period of Worldwide Harmonized Light Vehicles Test Cycle(WLTC).As a result,the consistency of the test evaluation results in the RDE tests can be improved by the proposed method. |